Just a list of bugs which should be fixed with the tree combiner.
Confirmed.
Mine, I will attach a patch here soon.
Created attachment 6325 [details] the new pass Right now this pass is very basic but can be improved. If someone would take a look at it I would be very happy. There still needs to be some comments added but it works and it solves some of the dependent bugs.
Will submit a new version which is much better of this version soon.
Patch posted here: <http://gcc.gnu.org/ml/gcc-patches/2005-05/msg02951.html> but it has a couple of regressions. First PR 21850. Second a latent bug in VRP saying the CFG will always be cleaned which is not true.
A number of xfailed tests in the tree-ssa test suite (e.g. bool-3.c) should start working if you add a tree combiner.
I don't have time for this bug.
tree-ssa-forwprop is becoming one but slowly. In the next few months, I am going to take what is done in fold-const and port them to forwprop. I already posted one of them: http://gcc.gnu.org/ml/gcc-patches/2012-01/msg00945.html . I will try to work on the rest and more.
This is fixed in a much better way by the following patch: https://gcc.gnu.org/git/?p=gcc.git;a=commit;h=2165588a33dc7d63840774b2aac61be999ef17ae